The question as used in the run (Polish)#
Jak wygląda doręczenia pism sądowych osobie fizycznej przez portal ?
Original wording, exactly as the lawyer typed it into the chat (Polish); typos kept.
The trap
A reversed legal-change trap. Until 31.05.2026 service through the portal concerned only professional legal representatives and a few public bodies. The answer "a natural person is not served through the portal" is outdated on the legal-state date. The question measures whether the system knows the amendment of 5.08.2025 (Journal of Laws 2025, item 1172).
The correct answer according to the key#
Core answer. Since 1.06.2026 the court serves documents through the information portal also on a party who chose this form of service (art. 131¹ᵃ § 1 of the Code of Civil Procedure). The party files the declaration of choice through the portal, with a qualified, trusted or personal signature (§ 1¹).
Required bases. art. 131¹ᵃ § 1 of the Code of Civil Procedure (wording from 1.06.2026); art. 131¹ᵃ § 1¹; art. 131¹ᵃ § 2.
Expected calibration: categorical. The B2 point (form of the conclusion) goes to an answer that states the conclusion unequivocally; blurring a clear thesis without saying what it would depend on costs the point.
The full key (rubric, penalties, quotation patterns from the source, known rulings) is in the repository in keys/P05_portal_doreczenia_osoba_fizyczna.yaml. Every legal basis was checked in the official text as of the legal-state date. Rules of the key.
Results of the eight models#
| Model | Condition | Block A % | Result | Card |
|---|---|---|---|---|
| GPT-6.1 Sol | W0 | 22 | – | card |
| W1 | 100 | ✓ | card | |
| GPT-6 Astra | W0 | 22 | – | card |
| W1 | 100 | ✓ | card | |
| GPT-5.6 Terra | W0 | 0 | W | card |
| W1 | 100 | ✓ | card | |
| Claude Opus 5.5 | W0 | 56 | ST | card |
| W1 | 89 | W | card | |
| Claude Sonnet 5.5 | W0 | 56 | T | card |
| W1 | 78 | W | card | |
| Claude Haiku 5.5 | W0 | 22 | T | card |
| W1 | 0 | FT | card | |
| Gemini 3.8 Flash | W0 | 6 | TW | card |
| W1 | 100 | ✓ | card | |
| Gemini 4 Argon | W0 | 0 | TW | card |
| W1 | 83 | TW | card |
correctly groundedno hallucination; not correctly groundedhallucinatedprovider blockColumns: models in report order; rows: W0 no tools, W1 with web search.
✓ correctly grounded · – no hallucination, but the core is incomplete or no source was checked · T W F S error classes in a hallucinated answer (definitions) · block text blocked by the provider's filter, no label · block A = criteria checked against the key and the source (scoring criteria)
The numbers concern one answer of each model to this question in each condition. They say nothing about run-to-run variance or about other questions in this field.